You are on page 1of 2

Machine Learning (15CS73)

Question Bank

Module 2

1. Define approximate inductive bias of ID3. Explain Restriction and Preference Bias .
2. List the issues in Learning decision trees and explain how to incorporate continuous
valued attributes in decision tree
3. List the issues in Learning decision tree with respect to information gain and explain the
alternative measures for selecting attributes in a decision tree
4. Define Over fitting. Explain the reduced pruning
5. Explain the Rule Post- pruning with example
6. State Occam’s razor. Explain how to handle attributes with missing values and differing
cost

Module 3

7. Define ANN. Briefly describe the characteristics in problems appropriate for Neural
Network Learning
8. Define Perceptron and give a single perceptron representation? Show how the AND
function can be implemented using s single pereceptron.
9. With example explain the perceptron training rule
10. With example explain the gradient descent and delta rule
11. Design a two-input perceptron that implements the boolean function A Λ ¬ B. Design a
two-layer network of perceptron’s that implements A XOR B.
12. Write the Gradient descent algorithm for training linear unit. Also explain the features of
gradient Descent Algorithm
13. Explain the stochastic approximation of gradient Descent Algorithm.Also differentiate
between standard and stochastic gradient descent.
14. Explain the differentiable sigmoid threshold unit. How can momentum be added to the
weight update rule of Back propagation algorithm . Explain its advantage
15. Write the back propagation algorithm
16. Derive the Back propagation rule for output unit weights
17. Derive the Back propagation rule for Hidden unit weights
18. Explain the problem of convergence and local minima. Also give the heuristics to
alleviate the problem of local minima
19. Explain the different functions that can be represented by a Feed Forward Network.
20. Explain the various techniques to address the over fitting problem for back propagation
learning
21. With diagram explain the representation of Hidden Layer? What is the appropriate condition to
stop the weight update?
22.Briefly Explain Generalization, Over fitting and Stopping Criterion

You might also like